The Keep, Compress, or Cut Score for Each Job on a Field-Change Resume

Changing fields turns every past job into a judgement call. A role that would obviously belong on a same-field resume now has to earn its place against a field you have never worked in, and the usual advice - "keep 10 to 15 years," "list only relevant roles" - stops exactly where the hard cases start. This guide replaces the eyeball with a score. You rate each role on six fixed dimensions, read the total into one of three verdicts (keep in full, compress to a line, cut), and get a bullet budget out the other end. It is built for people changing field, returning after a break, or rebuilding after a layoff, and it is written so two people scoring the same history land in the same place.

What the score decides, and why eyeballing fails

The score answers one repeated question: for a resume aimed at a field you have not worked in, does this past job stay in full, shrink to a single line, or come off entirely. That is a different question from "is this job good" or "am I proud of it," and conflating them is where most field-change resumes go wrong.

Eyeballing fails for two specific cases that the rules of thumb cannot resolve. The first is the recent off-field job: it is close in time, so recency guidance says keep it, but it belongs to the field you are leaving. The second is the old near-perfect match: it falls outside the 10-to-15-year window, so horizon guidance says cut it, but it maps directly onto the target role. A single rule cannot serve both, because they pull in opposite directions. You need a model that scores recency and relevance as separate dimensions and lets one override the other on the merits.

There is a hard constraint the score has to respect. In the widely cited Ladders eye-tracking study, recruiters spend an average of six to seven seconds on the first scan, and they read in an F-pattern that makes the current role the most scrutinised section. Attention is positional. A high-relevance old role, scored correctly, still loses if it sits below a bullet-heavy off-field recent job. So the verdict cannot only decide inclusion - it has to control order and bullet count too.

The six dimensions, and what each one proves

Score every role on six dimensions, each from 0 to 2, for a total out of twelve. Each dimension answers a different question, and each has a tell for when it is lying to you.

• Recency - how close in time the role is. Proves the skill is current. It lies when a recent role is off-field: closeness in time makes a job feel load-bearing when its content is not.
• Target-field relevance - how directly the role's work maps to the field you are entering. Proves the transition is real, not aspirational. It lies when a role shares a job title with the target but not the actual tasks.
• Transferable-skill density - how many skills the target field wants that this role demonstrably built. Proves you can do the new work even without the new title. It lies when skills are named but not evidenced with a result.
• Seniority signal - how much scope, ownership, or progression the role shows. Proves you operate at the level you are applying to. It lies when a senior title masks a narrow, off-field remit.
• Gap-bridging value - how much the role keeps your timeline honest. Proves there is no unexplained hole. It lies by omission: a role scores high here precisely because cutting it would open a gap, independent of its content.
• Achievement strength - whether the role carries a strong, quantified result. Proves work ethic and competence in a way an employer reads across fields. A former sales associate applying to an admin role can include it on the strength of performance numbers alone.

The dimensions are deliberately independent. Relevance can outweigh recency, so you do not shortchange a relevant past role just because it is old, and you do not over-reward a recent role just because it is near the top of the page. Gap-bridging value is the one dimension that can carry a role on its own: a visible unexplained gap is a first-scan knockout signal, and no amount of relevance elsewhere buys it back.

The verdict table: reading a score into keep, compress, or cut

A total out of twelve maps to one of three verdicts, with two override rules that can move a role up a band. The bands are a starting point; the overrides handle the cases the raw number gets wrong.

Two overrides sit on top of the bands. First, the gap override: if cutting a role opens an unexplained multi-year hole, promote it from cut to compress regardless of its content score. If you held a previous unrelated job for an extended period, omitting it may result in a gap, and a gap is worse than an off-field line. Second, the achievement override: a role scoring 0-3 overall but carrying one strong quantified result may be promoted to compress and carry that single result, because that number reads as competence across fields.

The matrix is the whole reason the score exists. The two corners people get wrong are top-left (recent, off-field), which feels important because it is recent, and bottom-right (old, on-field), which feels expendable because it is old. Scoring the axes separately forces the right call on both.

The horizon: how far back to go before you score

Set a cut-off year first, then let the score pull individual roles back across it. For most job seekers the resume covers the last 10 to 15 years, which usually captures three to five roles and keeps the document focused. A significant career change is the single most common reason to cap history under 10 years, and age-discrimination avoidance is a named reason experts stop around 15.

There is a competing anchor worth knowing about. One approach ties the window to the years a posting requires: if a posting calls for five years, take the resume back to about five years and give basic details of every job in that span regardless of relevance; eight to ten years required means going back that far. Sources disagree on which anchor leads. ResumeBuilder-style guidance anchors to years-required first; ResumeWorded and similar sources say relevance can override recency and pull older roles back in. My reading: set the horizon as a default fence, not a hard wall, and let a high relevance score jump the fence for an individual role.

Re-entry is the situation that most often extends the window. Substantial skill-building during a break - bootcamps, certifications, portfolio projects - can align more closely with the target role than your earlier employment history does, and when it does, it leads. The horizon then stretches not to reach an old paid role but to hold the break itself as a dated, purposeful entry.

Score it, tag it, budget it: the procedure

This is the full sequence from raw history to a two-section body. Work it in order; the inventory has to exist before you cut anything, or you will drop roles by forgetting them rather than by decision.

Scoring six roles on six dimensions is roughly thirty minutes of honest work, and the honesty is the hard part - it is tempting to inflate relevance for the job you liked. If you would rather not eyeball transferable-skill density against a specific posting, Refolk writes your resume from your own history and tailors it to each posting, then scores how well you actually fit, which turns the relevance dimension into a number instead of a guess.

Here is a copy-paste rubric to run against a single role.

ROLE: [title, employer, dates]
Recency (how current): ___
Target-field relevance (task overlap, not title): ___
Transferable-skill density (skills the target wants, with evidence): ___
Seniority signal (scope, ownership, progression): ___
Gap-bridging value (does cutting this open a hole?): ___
Achievement strength (a strong quantified result?): ___
TOTAL / 12: ___
Base verdict: KEEP (8-12) / COMPRESS (4-7) / CUT (0-3)
Override applied? gap -> promote to COMPRESS / strong number -> promote to COMPRESS
Final verdict + bullet count: ___

The bullet budget: how many lines a kept role earns

A kept role earns bullets by verdict and by recency, and the most recent position always gets the most real estate. The counts below are documented ranges, not invention.

Two rules govern the budget. The ceiling on your most recent role is seven to eight bullets, and it is a ceiling, not a target; if you are stretching to eight everywhere, the F-pattern scan never reaches your older relevant roles. The floor is firmer: never leave a role with exactly one bullet, because one line reads as an omission, not a summary. If a role only justifies one bullet, either give it a title-only line with no bullets (the "Early Career Experience" convention handles this cleanly) or find a second real point.

This is where the two tables meet the six-second constraint. A role can score 9 out of 12 and still be an old role, and an old role gets few bullets even when kept. Keeping it in full means keeping the entry and the relevance, not handing it the bullet count of a current job. The budget protects the front-loading that the scan demands.

How supply density should set your cut aggression

How aggressively you cut off-field roles should depend on how crowded the target field is. In a dense field you are competing against tens of thousands of on-title incumbents, which raises the bar for any off-field role to keep its place. In a thinner field, a well-scored transferable role clears more easily, because there are fewer pure-pedigree candidates ahead of it.

The size of that effect is measurable. In Refolk's index of professional profiles, 63,122 US professionals hold a Data Analyst title against 15,405 in the UK - a supply gap of about 4.1x.

There is a second signal in the same data. Data Analyst made up about 84% of sampled top titles in the UK versus roughly 60% in the US, which implies US analyst pipelines pull from more varied prior titles. Where a field already absorbs entrants from many backgrounds, a well-scored transferable role is normal, not a liability, and you can keep it with more confidence. Where the title is "purer," your off-field roles have to work harder to justify staying, so cut more aggressively and lean on the strongest one or two.

To see who actually made your target transition and what their prior titles were, you can search the population directly instead of guessing which past field reads as credible.

How this goes wrong: failure modes and false positives

Most bad field-change resumes are not built from bad roles; they are built from good roles filed under the wrong verdict. These are the recurring errors and the check that catches each one.

• Recency beats relevance by mistake. A recent off-field job gets full bullets while an older near-perfect match is cut. The whole score exists to prevent this. Check: score relevance independently of recency, and do not shortchange a relevant past role just because it is old.
• Cutting a long unrelated role opens a gap. The resume looks tighter but now shows an unexplained multi-year hole, which is a first-scan knockout. Check: if you held an unrelated job for an extended time, compress it to a line instead of cutting it. The gap override exists for exactly this.
• Single-bullet compression backfires. Shrinking a role to one bullet reads as concealment, not economy. Check: compress to a title-only line or give two to three bullets, never exactly one.
• Reaching for a functional format to hide the pivot. A skills-first resume with no dated timeline looks strong but recruiters view it with suspicion, assuming you are hiding gaps or a lack of experience. Check: use a hybrid that keeps a dated work history, and let the Relevant/Additional split carry the emphasis.
• Keyword stuffing to game the ATS. The ATS ranks on keywords, which tempts people to overload them or hide them in white font. Check: an ATS rarely auto-rejects; it ranks. Stuffing gets flagged and hidden text gets caught. Earn the keywords with real content.
• Under-weighting volunteer and project work because it was unpaid. The strongest, most recent, most relevant work ends up at the bottom of the page because it did not come with a salary. Check: treat volunteer, freelance, and part-time work with the same rigor as paid roles and place it by relevance.

A gap is louder than an off-field line, so when in doubt compress the role rather than cut it.

There is one tooling quirk worth naming if you go looking in a profile index yourself. Filtering for a base title plus a separate seniority flag can return zero results, because seniority often lives inside the exact title string - "Senior Data Analyst" - rather than in its own field. Query the full senior title, not the base title with a seniority filter bolted on.

Keep it current: a final check before you send

Before you call the resume done, verify the score actually produced the document, not the other way around. The checklist below is the last pass; run it against the finished file, not your intentions.

Re-score when the target changes, not on a calendar. The same history produces different verdicts against different fields, because relevance and transferable-skill density are scored against a specific target, and the supply density that sets your cut aggression differs by role and market. When you switch which posting you are aiming at, the horizon can move, an off-field role can flip from cut to keep, and the bullet budget shifts with it. Rescoring is cheaper than it looks once the score sheet exists, and it is the difference between a resume tuned to a field and a generic history hoping to pass. If you would rather not rebuild it by hand each time, Refolk tailors the resume to every posting and scores the fit, so the re-score happens per application instead of once.